The present invention relates to the production of covalently closed circular (ccc) recombinant DNA molecules. Such molecules are useful in biotechnology, transgenic organisms, gene therapy, therapeutic vaccination, agriculture and DNA vaccines.
With the invention in mind, a search of the prior art was conducted. E. coli plasmids have long been the single most important source of recombinant DNA molecules used by researchers and by industry. Today, plasmid DNA is becoming increasingly important as the next generation of biotechnology products (gene medicines and DNA vaccines) make their way into clinical trials, and eventually into the pharmaceutical marketplace. Plasmid DNA vaccines may find application as preventive vaccines for viral, bacterial, or parasitic diseases; immunizing agents for the preparation of hyper immune globulin products; therapeutic vaccines for infectious diseases; or as cancer vaccines.
Fermentation
Vector Backbones
Therapeutic plasmids typically contain a pMBI, ColE1 or pBR322 derived replication origin. Common high copy number derivatives have mutations affecting copy number regulation, such as rop (Repressor of primer gene) deletion, with a second site mutation that increases copy number (e.g. pMB1 pU CG to A point mutation, or ColE1 pMMI). Higher temperature (42° C.) can be employed to induce selective plasmid amplification with pUC and pMMI replication origins.
Nature Technology Corporation Fed-Batch Process:
Carnes, A E., Williams, J A 2006 World Patent Application WO2006023546 discloses methods for fed-batch fermentation, in which plasmid-containing E. coli cells are grown at a reduced temperature during part of the fed-batch phase, during which growth rate is restricted, followed by a temperature up-shift and continued growth at elevated temperature in order to accumulate plasmid; the temperature shift at restricted growth rate improves yield and purity of plasmid.
This process takes advantage of the temperature sensitivity of high copy number plasmids. In the preferred process, the initial temperature setpoint is 30° C., at which the plasmid is maintained stably at low levels while biomass can accumulate efficiently. During this period, the specific growth rate is controlled at approximately μ=0.12 h−1 by an exponential feeding strategy. Induction of plasmid accumulation is performed when the cell density is in the range of 25-60 OD600 by shifting the temperature to 42° C. and continued exponential nutrient feeding for up to 15 hours.
Plasmid yields prior to the temperature shift remain low. The specific plasmid yields after temperature shift are very high. Interestingly, after the temperature shift, the cells are able to tolerate significantly higher quantities of plasmid than cells grown at a constant temperature of 37° C. with the same media and feeding strategy.
The examples in the patent report yields up to 1.1 g/L achieved when the disclosed process was used with a temperature shift from 30° C. to 42° C. The preferred process of Carnes and Williams, Supra, 2006 is also described in Carnes, A E, Hodgson C P, Williams J A. 2006 Biotechnol Appl Biochem 45:155-66 where volumetric yields of 1.5-2.1 g/L, and specific yields as high as 43 mg plasmid/g dry cell weight (DCW) are reported. The plasmid DNA produced with the process is of a high quality, being 96% supercoiled or greater with no detectable deletion or other rearrangement. The method is simple, can be used with multiple pUC based backbones, and does not require prescreening of individual colonies for high producing cell lines. A key advantage of the inducible fed-batch process is that amplification of plasmid copy number after suitable biomass accumulation helps preserve quality and stabilize toxic plasmids, while maximizing yield. This is because selection pressure at the cellular level is reduced during the biomass accumulation phase by minimizing the growth rate difference between monomer containing cells and dimer, or rearranged plasmid or plasmid-free cells.
Fermentation Summary
High specific yields are very desirable since increased plasmid yield per gram of bacteria, or increased plasmid relative to genomic DNA (utilizing the Carnes and Williams, Supra, 2006 process, up to 75% of the total DNA in the cell at harvest is plasmid DNA) lead directly to higher final product purities. Further improvements in yield or increases in the percentage plasmid per total DNA would further decrease production cost, improve purity and simplify removal of genomic DNA (gDNA). Other fermentation processes for plasmid production are reviewed in Carnes et al., Supra, 2006, and Carnes A. E. BioProcess Intl 2005; 3:36-44, and are included herein by reference.
Cell Disruption—Plasmid Release
The E. coli biomass generated by a fermentation process must be lysed to release the plasmid DNA (pDNA). Cell disruption methods fall into two main categories:
Physico-mechanicalChemicalliquid sheardetergentssolid shearosmotic shockagitation with abrasivesalkali treatmentfreeze-thawingenzyme treatmentultrasonicationheat
The cell disruption method for plasmid isolation must be chosen such that minimal damage is inflicted on the pDNA product, and in most cases, it is also desired to avoid shearing of the host cell gDNA into smaller fragments that are more difficult to separate from pDNA. Thus, the methods available for plasmid purification are more limited compared to the harsher methods that are often used for purifying smaller molecules such as proteins. Ideally the method releases a high yield of intact plasmid, while limiting release of difficult to remove impurities such as gDNA.
Most of the above methods have been applied, either alone or combined, to pDNA purification. To date, the two most commonly used methods for pDNA recovery are alkaline lysis and heat lysis; additionally, detergents and enzyme (i.e. lysozyme) treatment are often used to aid heat lysis and other methods.
Alkaline Lysis
The standard alkaline lysis method of Birnboim H C, Doly J. 1979 Nucleic Acids Res. 7:1513-23 is well known and is widely used without restriction in molecular biology laboratories. Generally, a lysis time of five minutes has been used with the standard alkaline lysis method; longer times have been known to cause irreversible denaturation of pDNA. A study (Ciccolini L A, Shamlou P A, Titchener-Hooker N J, Ward J M, Dunnill P. 1998 Biotechol Bioeng. 60: 768-770) on the time course of standard alkaline lysis was performed by measuring viscosity of a lysis mixture as a function of time and by performing cell counts over a range of lysis times. The results indicate that for E. coli DH5α, complete cell lysis occurs after about 40 sec and complete denaturation of gDNA takes 80-120 sec after mixing with the lysis buffer. Longer reaction times were reported to lead to shear degradation of gDNA.
Thatcher D R, Hitchcock A, Hanak J A J, Varley D. 2003 U.S. Pat. No. 6,503,738 describe a method to determine the optimum lysis pH value, which is about 0.2 pH below the “irreversible alkaline denaturation value”, defined as “the pH value at which no more than about 50% of the alkaline denatured pDNA fails to renature as determined by standard agarose gel electrophoresis”. The optimum lysis value can be different for various plasmid/host strain combinations.
The patent landscape includes various methods and devices aimed at performing alkaline lysis at large scale. Insufficient mixing will result in local pH extremes, causing irreversibly denatured plasmid. Mixing that is too aggressive can damage the pDNA and fragment gDNA. At the laboratory scale, mixing is performed gently by hand. Hand mixing at larger scales is not possible because of the large volumes and lack of reproducibility from person to person. Thus, batch mixing in a mechanically agitated vessel is often used, but the viscous, non-Newtonian properties of the lysis mixture require some consideration. Nienow A W, Hitchcock A G, Riley G L. 2003 U.S. Pat. No. 6,395,516 discloses a specialized vessel design for mixing cell lysate that utilizes baffles, low power number impellers, feed lines, and monitoring the degree of lysis by measuring viscosity.
Continuous flow through devices have been employed as alternatives to the challenge of achieving complete, but gentle mixing of large lysis volumes in stirred tanks, and are perhaps more easily implemented in facilities that do not already contain specialized batch mixing equipment. Additionally, the lysis reaction time can be closely controlled by the residence time of tubing or pipe (e.g. as in Wan N C, McNeilly D S, Christopher C W. 1998 U.S. Pat. No. 5,837,529; Chevalier, M 2003 U.S. Pat. No. 6,664,049; Detraz N J F, Rigaut G. 2006 World Patent Application WO2006060282), or of the holding vessel (e.g. as in Brooks R C. 2004 U.S. Pat. No. 6,699,706) before the neutralization step.
Inline static mixers (motionless mixers) have long been used in industry and more recently have been applied for cell lysis. A static mixer is a cylindrical tube containing stationary mixing elements. The mixing elements are shaped and positioned to combine materials as they flow through the mixer. Wan et al, Supra 1998 describes the use of static mixers to achieve gentle mixing of a cell suspension with a lysis solution. Mixing of the cell suspension stream with the lysis buffer stream is completed rapidly and the degree of mixing and lysis time can be adjusted by the number of mixing elements, flowrate, and length. Neutralization may occur in a second static mixer.
Chevalier, Supra, 2003 claims mixing methods that use only tubing, without the static mixers; instead, smaller diameter tubing is used and flowrates are adjusted to cause homogeneous mixing for the desired contact time.
Detraz and Rigaut, Supra, 2006 discloses flow-through mixing devices consisting of a conduit through which the lysis solution flows, and an inlet, such as a nozzle, into the conduit in which the cell suspension is injected in either a counter-flow or co-flow direction.
Brooks, Supra, 2004 claims the use of fluidic vortex mixers for continuous flow-through lysis and neutralization. A vortex mixer is described by the patent as a cylindrical chamber with an axial outlet at the center of one end wall with two tangential inlets along the periphery. The dimensions of the mixer and the flowrates used are chosen so that the residence time in the mixers is much less that the time required for lysis (about 0.01-0.1 sec) so that the cell suspension and lysis solution are mixed completely. The cells can then react with the lysis solution after exiting the vortex mixer. In an example, the cells and lysis solution are mixed in a first vortex mixer, flow into a tank for completion of lysis, and then the mixture flows through an outlet of the tank where it is mixed with neutralization buffer in a second vortex mixer.
Blanche F, Couder M, Maestrali N, Gaillac D, Guillemin T. 2005 World Patent Application WO2005026331 discloses continuous alkaline lysis through the use of T tubes with lengths of turbulent flow (achieved by small diameter tubing) to rapidly mix the cell suspension and lysis solution, followed by a length of laminar flow (in larger diameter tubing) for incubation and time for lysis and denaturation without substantial agitation which would damage the plasmid and fragment gDNA. Neutralization solution may then be introduced continuously in a second T tube.
The above flow-through mixing devices enable low shear mixing. It has been generally recognized that shear forces created by mixing too intensely may cause damage to pDNA and fragmentation of gDNA, leading to co-purification of gDNA with pDNA (Horn N A, Meek J A, Budahazi G, Marquet M. 1995 Hum. Gene Ther. 6: 565-573).
Alternative Lysis Methods
Heat Lysis
Plasmid isolation using heat lysis was first reported by Holmes D S, Quigley M. 1981 Anal Biochem 114:193-7, and is perhaps the most widely used method after alkaline lysis.
Merck has developed and patented processes to adapt heat lysis to large scale processing. In Lee, A L, Sagar, S. 2001 U.S. Pat. No. 6,197,553, a bacterial suspension in modified STET buffer (e.g. 50 mM Tris, 50-100 mM EDTA, 8% sucrose, 2% Triton X-100, pH 8.0-8.5) with a density of about 30 OD600 is pumped through a heat exchanger at such a rate that the suspension exits with a temperature of 70-100° C., resulting in lysis. The lysate is then centrifuged to pellet large cell debris, protein, and gDNA, leaving RNA and plasmid in solution. The optional use of lysozyme is reported to increase the plasmid concentration in the lysate by 4-5 times. It was also determined that the formation of undesirable open circle plasmid by endogenous DNase during this lysis process could be reduced by increasing the EDTA concentration from 50 mM to 100 mM. They report higher plasmid recovery than by chemical lyses.
A similar process is described by Zhu K, Jin H, Ma Y, Ren Z, Xiao C, He Z, Zhang F, Zhu Q, Wang B. 2005 J Biotechnol. 118: 257-264, which reports to have made improvements on the heat lysis methods of Holmes and Quigley, Supra, 1981, and Lee and Sagar, Supra, 2001. In this protocol, cell paste is resuspended with 10 mM Tris, 50 mM EDTA, pH 8.0 to a density of 100 OD600 and treated with 0.1M NaCl, 2% Triton X-100, and lysozyme at 37° C. for 20 minutes. The cell suspension is then pumped through a copper coil immersed in a 70-80° C. water bath with a residence time of 20 sec, then it enters another copper coil immersed in an ice bath.
Mechanical Disruption
Generally, mechanical disruption of bacteria (e.g. french press, sonication, homogenization, nebulization) for plasmid isolation is seen as unfeasible due to the damage it would cause to the DNA. Jem K J. 2002 U.S. Pat. No. 6,455,287 reports that sonication, nebulization, and Gaulin Mill homogenization resulted in almost complete destruction of pDNA. However, disruption with a bead mill device under optimized conditions resulted in over 90% of the plasmid solubilized without destruction. They also report that an impinging-jet homogenizer released up to 50% of the pDNA intact.
Another method used to overcome destruction of DNA during mechanical disruption is the use of DNA compaction agents. Wilson R C, Murphy J C. 2002 US Patent Application US2002197637 disclose the use of polycationic compaction agents (e.g. polylysine, spermine, spermidine) to protect DNA from shear damage during mechanical lysis. The compaction agents cause the DNA to be pelleted with the insoluble cell debris. The pellet is washed, and the pDNA is resolubilized to give an enriched solution. The use of compaction agents also results in reduced lysate viscosity.
Lysozyme Lysis
A process developed by Merck (Boyd D B, Kristopeit A J, Lander R J, Murphy J C, Winters M A. 2006 World Patent Application WO2006083721) describes a STET/lysozyme lysis performed at 20° C. or 37° C., preferably with an additional alkaline pH shift to denature gDNA. While the process retains the pH shifting of alkaline lysis, shear forces are reduced by performing the shift after lysis. Therefore this process does eliminate many of the difficult processing and equipment needs of alkaline or heat lysis. The limitation of this reduced temperature lysis method is the need for large amounts of recombinant lysozyme.
Autolysis
Autolytic strains using phage T4 lysis proteins have been patented for protein production as in Leung W S, Swartz J R. 2001 U.S. Pat. No. 6,258,560. In this system, lysozyme (endolysin) is expressed by the cell in the cytoplasm and released to the periplasm at the desired time by co-expression of a holin (membrane spanning peptide or protein) that creates a channel, allowing leakage of lysozyme from the cytoplasm to the periplasm. Other autolytic E. coli strains that are described in Jia X, Kostal J, Claypool J A. 2006 US Patent Application US20060040393 contain the bacteriophage λR lytic endolysin gene. The endolysin is induced by arabinose, which then causes the E. coli to be lysed after a freeze-thaw cycle.
Autolysis conditions, as opposed to alkaline or heat lysis, do not selectively denature gDNA. The product of lysis is very viscous due to high levels of residual gDNA, creating processing problems. For protein production, non specific nucleases (e.g., Benzonase®) are added, or expressed periplasmically in the strain [e.g., endA or Staphylococcus nuclease (Cooke G D, Cranenburgh R M, Hanak J A J, Ward J M. 2003 J Biotechnology 101: 229-239)] to reduce viscosity after cell lysis. Such systems could not be utilized for plasmid production, since the plasmid would be degraded or damaged by the nuclease. While autolysis is not an essential design improvement for protein production (since cell lysis is performed at high density using generally available equipment) it has tremendous potential for plasmid purification since alkaline or heat lysis steps are key bottlenecks in plasmid processing.
Bacteriophage T5 exonuclease is an ideal DNase to use in plasmid processing. T5 exonuclease does not digest supercoiled plasmid, but is able to digest linear single- and double-stranded DNA (ssDNA, dsDNA). It will also digest DNA with denaturation loops, such as ‘ghost’ or ‘shadow band’ DNA, which often retains biological activity and is refractile to restriction enzyme digestion. Williams J A, Hodgson C P. 2006 World Patent Application WO2006026125 describe plasmid purification using E. coli strains expressing plasmid-safe nuclease (chimeric ribonuclease-T5 exonuclease genes) in combination with endolysin/holin pairs for autolysis.
Cell Disruption Summary
While the basic methods for obtaining plasmids (by bacterial fermentation), and for their purification (e.g., by heat or alkaline lysis) are well-known, and large scale manufacturing methods have been developed, these processes are problematic for transfer to new facilities due to specialized equipment needs, scaling issues, and tremendous lysis volumes. As well, they add excessive additional cost to the production of plasmids through reduced capacity, increased wastestreams, and expensive equipment and reagents. These limitations place a cost burden on commercialization of pDNA production processes. A new technology is needed to eliminate this critical processing bottleneck.